{"id":"f8da8b26-3522-400f-bc78-9f12cfba99ce","arxiv_id":"2504.14950","paper_version":3,"verdict":"UNVERDICTED","confidence":"UNKNOWN","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A thin-film periodically poled lithium niobate ridge waveguide with 0.106 dB/cm propagation loss and 805%/W second-harmonic conversion efficiency was fabricated by chemo-mechanical polishing and post-etch poling.","lead":"Researchers made a 7-millimeter lithium niobate waveguide with record-low light loss of 0.106 dB/cm by polishing it smooth before adding a periodic electric-field pattern. The same chip converted laser light to its second harmonic at near-record efficiency, suggesting cheaper and more scalable nonlinear photonic circuits.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The record-low-loss claim rests on a ring-Q measurement, not on the 7-mm straight waveguide used for SHG; direct straight-waveguide loss data are needed to support the central claim.","rationale":"The paper's novelty and impact rest on the record-low propagation-loss claim. The only loss evidence is an intrinsic-Q measurement on a 2-mm racetrack ring, while the nonlinear result is obtained in a 7-mm straight ridge waveguide. That structural gap is load-bearing because the low loss is used to explain the high conversion efficiency, and because the manuscript supplies no repeated devices, raw transmission curves, or error bars. The internal formulas (poling period, loss-from-Q, normalized efficiency from slope and length) are consistent, and I found no contradiction in the SHG analysis. The reader's weakest_assumption identified the same transferability concern, so I agree. I recommend UNCHANGED: the report remains UNVERDICTED pending a direct straight-waveguide loss measurement; no verdict move is needed based on this pass.","tokens_in":6749,"tokens_out":9170,"duration_ms":87318,"concrete_test":"Fabricate at least three straight PPLNOI ridge waveguides with identical cross-section but different lengths (e.g., 3, 5, and 7 mm) on the same chip, using the same PLACE flow and high-voltage poling as the reported device. Measure fiber-to-fiber transmission at 1550 nm for each, subtract the calibrated facet coupling losses, and extract the propagation loss from the length-dependent slope. If the derived loss is below about 0.15 dB/cm, the ring-Q estimate transfers; if it exceeds 0.2 dB/cm, the record-low loss claim and the efficiency interpretation should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that PLACE waveguides reach record-low 0.106 dB/cm is supported only indirectly. The paper states that 'the propagation loss was measured to be 0.106 dB/cm ... by measuring the intrinsic Q factor of a racetrack PPLNOI microring with a physical length of 2 mm, which was 3.1×10^6.' The SHG demonstration, however, is on a 7-mm straight ridge waveguide. Using the ring loss as the loss of that straight device assumes transferability of sidewall roughness, domain duty cycle, and poling quality between the two structures. The ring contains bends and a closed geometry; even if bend loss would usually make the ring value an upper bound for a straight guide, the two devices are not shown to be co-fabricated on the same chip, and no repeated devices or error bars are reported. If the straight waveguide actually has higher loss, the 'record-low' statement is unsupported and the net SHG efficiency interpretation changes. This assumption is the least secure link in the paper's central advance.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The manuscript reports the fabrication of a 7-mm-long periodically poled lithium niobate on insulator (PPLNOI) ridge waveguide using photolithography-assisted chemo-mechanical etching (PLACE) followed by high-voltage pulse poling. The authors measure an average surface roughness of 0.27 nm, extract a propagation loss of 0.106 dB/cm from the intrinsic Q of a 2-mm racetrack microring, and demonstrate second-harmonic generation with a normalized efficiency of 1643%·W⁻¹·cm⁻² and a slope of 805%/W at room temperature, increasing to 1742%·W⁻¹·cm⁻² at 59 °C. The central claims are a record-low propagation loss in PPLNOI waveguides and near-record SHG conversion efficiency at low fabrication cost.","tokens_in":6866,"tokens_out":8680,"duration_ms":72664,"significance":"If the loss and efficiency figures are robust, the work is significant: it provides a low-cost, lithography-based alternative to electron-beam lithography for PPLNOI waveguides, with ultra-smooth sidewalls that mitigate scattering loss. The internal consistency between the QPM period calculation and the SHG wavelength, as well as the agreement between the normalized efficiency and the slope (805%/W over 0.7 cm yields 1643%/W/cm²), support the plausibility of the reported conversion performance. The paper, however, does not supply uncertainty estimates or repeated-device statistics, and the loss measurement is not performed on the straight waveguide used for SHG.","major_comments":[{"comment":"The propagation loss of 0.106 dB/cm is extracted from the intrinsic Q factor of a 2-mm racetrack PPLNOI microring, but the SHG device is a 7-mm straight ridge waveguide. The manuscript does not state explicitly that the racetrack and the straight waveguide were co-fabricated on the same chip with identical cross-section and poling parameters, nor does it show the racetrack structure. Because the record-low-loss claim is a central advance, a direct measurement on the straight waveguide (e.g., cut-back or out-scattering) or a clear demonstration of co-fabrication with an upper-bound argument for bend loss is needed. Additionally, no error bars or multiple devices are reported, so the statistical basis for the 'record-low' claim is not established.","section":"Fabrication and characterization (Fig. 2 and propagation loss paragraph)"},{"comment":"The comparison to the record conversion efficiency (814%/W, ref. 28) is made using a non-normalized slope (805%/W) without specifying the length of the reference device. The authors also report a normalized efficiency (1643%·W⁻¹·cm⁻²), which is substantially lower than some published normalized values (e.g., 4600%·W⁻¹·cm⁻² in ref. 24). The benchmark should be stated in consistent units (slope versus normalized efficiency) and with the relevant lengths and coupling-loss conventions for the reference, so that the claim of being 'close to the best' is quantitatively justified.","section":"Results, SHG efficiency comparison (Fig. 4(e) and ref. 28)"}],"minor_comments":[{"comment":"The term 'PLCAE' appears in the fabrication description and conclusion, inconsistent with the acronym 'PLACE' used elsewhere; also the title contains a stray space in 'waveguide s'.","section":"Throughout (title, abstract, fabrication steps)"},{"comment":"'closed to the best' should be 'close to the best' (abstract, main text and conclusion).","section":"Abstract and text"},{"comment":"The designed poling period of 3.62 μm is calculated for a 1550 nm pump, but the measured phase-matching peak at room temperature is at 1556.56 nm; a brief comment on this 6.5 nm discrepancy (e.g., fabrication tolerances or index-model accuracy) would clarify the reliability of the QPM design.","section":"Fabrication and characterization (poling period calculation)"},{"comment":"The calibration of the ~11 dB/facet and ~24 dB/facet coupling losses is described only in one sentence; a short explanation of how these values are de-embedded from the linear transmission measurement would help the reader understand the quoted on-chip efficiencies.","section":"Experimental setup paragraph"},{"comment":"The units of normalized efficiency are typeset inconsistently ('%W-1·cm-2' vs '%·W⁻¹·cm⁻²') across the abstract, main text, and figure captions.","section":"Figures and typography"},{"comment":"Reference 27 contains a typo in the title ('Litium'); also, the temperature values in Fig. 4(b)–(d) are given as 24.8 °C, 44.5 °C, and 59.0 °C, but the text refers to the first as 'room temperature' without specifying whether this was an actively controlled temperature.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a compact Letter with an appealing fabrication approach. The main risk is the over-claim of a record-low loss based on a single ring measurement without uncertainty analysis or co-fabrication details. The efficiency benchmarking also needs to be made consistent in units and reference-device dimensions. If the authors can supply co-fabrication evidence, repeated measurements, and corrected benchmarking, the paper could be acceptable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nThis is a solid engineering result. The authors apply their PLACE fabrication technique (femtosecond laser photolithography plus chemo-mechanical polishing) to periodically poled LNOI, with poling done after etching. They report 0.27 nm sidewall roughness and a propagation loss of 0.106 dB/cm extracted from a racetrack microring’s intrinsic Q of 3.1×10^6. That is the lowest loss I have seen for PPLNOI waveguides, and the SHG numbers—1643%/W/cm^2 normalized, 805%/W slope—are close to the best e-beam/dry-etched devices. The QPM period calculation is standard, the loss-from-Q arithmetic is consistent, and the normalized efficiency matches the slope over 0.7 cm. The process is cheaper than e-beam lithography, which matters for scalable nonlinear photonics.\n\nI agree with the stress-test note: the central claim rests on a ring measurement, not on the 7-mm straight waveguide used for SHG. The ring contains bends and may not perfectly represent the straight guide; the paper does not state that they are co-fabricated on the same chip, reports no repeated devices, and gives no error bars. That is a legitimate soft spot, but not fatal. The ring and straight guide almost certainly share the same cross-section and fabrication run, so the value is probably representative. Still, a direct straight-waveguide loss measurement (cut-back or Fabry-Perot) would make the claim much stronger. The efficiency definition is also slightly ambiguous: the phrase “without deducting propagation loss” and the large coupling losses (11 and 24 dB/facet) leave it unclear whether quoted powers are on-chip or fiber-referenced. Clarifying this would help readers compare with other work.\n\nThese are presentation and statistics issues, not flaws in the core approach. Post-etch poling works, the sidewalls stay smooth, and the device performs well. The paper deserves a serious referee, not a desk reject. I would ask for error bars, a straight-waveguide loss measurement, and an explicit statement of how efficiency was normalized. With those additions, this would be a reliable reference data point. As is, it is a credible but slightly under-supported record claim.\n\nFor a reading group, it is worth a quick look if people care about LNOI fabrication, but it is not conceptually deep. I would likely cite it as a fabrication data point if I were writing about low-loss PPLNOI.\n\nBest,\n[Your name]","headline":"A credible fabrication advance—record-low loss for PPLNOI waveguides via PLACE and post-etch poling—but the headline loss number comes from a ring measurement, not the actual SHG device, and error bars are missing.","tokens_in":7535,"tokens_out":2804,"would_cite":true,"duration_ms":27217,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["42.65.Ky","42.82.Et"],"model":"deepseek-v4-flash","headline":"The paper claims that periodically poled lithium niobate ridge waveguides made by femtosecond-laser photolithography-assisted chemo-mechanical etching reach a record-low propagation loss of 0.106 dB/cm, with near-record second-harmonic…","keywords":["periodically poled lithium niobate","thin-film lithium niobate","PPLNOI waveguides","femtosecond laser photolithography","chemomechanical etching","second-harmonic generation","propagation loss","quasi-phase matching"],"falsifier":"Measure the straight-waveguide propagation loss directly by cut-back: compare end-to-end transmission through PPLNOI ridge waveguides of different lengths and see whether the extracted loss matches 0.106 dB/cm, and repeat the SHG slope measurement with independently calibrated facet coupling.","tokens_in":6481,"feed_emoji":"🔬","tokens_out":8369,"duration_ms":68676,"temperature":0.7,"pith_summary":"The paper claims that periodically poled lithium niobate on insulator (PPLNOI) ridge waveguides can be made with record-low loss by reversing the usual fabrication order: first cut ultra-smooth ridges with chemo-mechanical etching through a femtosecond-laser-written mask, then invert the ferroelectric domains with high-voltage pulses. The resulting 7-mm waveguide shows a propagation loss of 0.106 dB/cm, about half the best prior value for such devices, and converts a telecom pump into second-harmonic light at 805% per watt, close to the record set by electron-beam-lithographed waveguides. If true, this means high-efficiency nonlinear photonics could be fabricated at lower cost and larger scale than the current e-beam-plus-dry-etch route. A sympathetic reader would take the central claim to be that fabrication order and surface smoothness, not material quality, are the main barriers to low-loss PPLNOI devices.","feed_headline":"Laser-written masks cut lithium niobate waveguide loss to 0.106 dB/cm","feed_subtitle":"Poling after chemo-mechanical etching yields near-record second-harmonic conversion without electron-beam lithography.","key_machinery":"The load-bearing mechanism is PLACE, photolithography-assisted chemo-mechanical etching, followed by post-etch high-voltage pulse poling. A femtosecond laser ablates a chromium mask on the lithium niobate film, chemo-mechanical polishing transfers the ridge pattern into the film with sub-nanometer smoothness, and comb-shaped electrodes then apply 400-V pulses to invert the domains beneath the ridge. The poling period is set by the quasi-phase-matching condition $\\Lambda = \\lambda_{2\\omega}/(n_{2\\omega}-n_\\omega)$, giving 3.62 µm for a 1550-nm pump.","core_discovery":"On the paper's own terms, the discovery is that doing chemo-mechanical etching before ferroelectric poling eliminates the two known loss sources in PPLNOI: rough sidewalls and etch-rate differences between oppositely poled domains. The PLACE process, followed by 400-V pulse poling through comb electrodes, produces sidewalls with 0.27 nm average roughness and a single-mode 7-mm ridge waveguide whose loss is measured at 0.106 dB/cm in the telecom band. That waveguide then delivers quasi-phase-matched second-harmonic generation with normalized efficiency 1643% W⁻¹ cm⁻² and a slope efficiency of 805%/W at 24.8 °C, improving to 1742% W⁻¹ cm⁻² at 59 °C, with absolute conversion of 15.8% at 21.6 mW pump power.","pith_inferences":["If the ring-derived loss transfers to straight guides, this fabrication route should also benefit quantum photon-pair sources, where on-chip loss directly suppresses pair-generation brightness.","A cut-back measurement on straight waveguides would be a stricter test than ring-Q loss, since bends and coupling can either mask sidewall scattering or add loss not present in the SHG device.","Combining PLACE with duty-cycle engineering or chirped poling could plausibly push normalized efficiency beyond the current benchmark, since the method already removes the loss that limits net conversion.","The same laser-mask-plus-CMP sequence should transfer to other ferroelectric thin films such as lithium tantalate, where dry-etch roughness is likewise a bottleneck."],"forward_implications":["Over the 7-mm device length, 0.106 dB/cm means roughly 1.7% total propagation loss, so most pump power stays in the waveguide and longer cascaded devices become practical.","Because poling occurs after etching, the method avoids domain-selective etch rates and can be applied to any pre-etched LNOI ridge without extra processing.","The room-temperature SHG slope of 805%/W approaches the 814%/W record from e-beam/dry-etched PPLNOI, indicating comparable nonlinear performance at lower fabrication cost.","Raising the device temperature to 59 °C shifts the phase-matching wavelength from 1556.56 nm to 1561.51 nm and increases normalized efficiency to 1742% W⁻¹ cm⁻²."],"supporting_citations":[{"why":"Provides the previous best PPLNOI propagation-loss baseline of 0.23 dB/cm before poling and 0.25 dB/cm after, which this work claims to improve twofold.","marker":"[25]"},{"why":"Introduces the PLACE process showing sub-nanometer roughness in LNOI waveguides, the fabrication method this work adapts.","marker":"[27]"},{"why":"Reports the 814%/W SHG conversion efficiency in e-beam/dry-etched PPLNOI, the record this work compares against.","marker":"[28]"},{"why":"Demonstrates ultrahigh-efficiency wavelength conversion in nanophotonic PPLN waveguides, establishing the efficiency class this work aims to approach.","marker":"[21]"}],"fun_headline_variants":["Record-low PPLN loss: 0.106 dB/cm via chemo-mechanical etching","Etch-before-pole yields record 0.106 dB/cm in PPLN waveguides","0.106 dB/cm: new benchmark for PPLN waveguide loss","Poling after etch sets PPLN loss record at 0.106 dB/cm","Lowest PPLN loss yet: 0.106 dB/cm via chemo-mechanical route"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the 0.106 dB/cm propagation loss derived from the intrinsic Q factor of a 2-mm racetrack microring is representative of the 7-mm straight ridge waveguide actually used for second-harmonic generation.","fun_headline_variants_meta":{"raw":{"variants":["Record-low PPLN loss: 0.106 dB/cm via chemo-mechanical etching","Etch-before-pole yields record 0.106 dB/cm in PPLN waveguides","0.106 dB/cm: new benchmark for PPLN waveguide loss","Poling after etch sets PPLN loss record at 0.106 dB/cm","Lowest PPLN loss yet: 0.106 dB/cm via chemo-mechanical route"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000964,"raw_usage":{"total_tokens":4141,"prompt_tokens":1020,"completion_tokens":3121,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":636,"completion_tokens_details":{"reasoning_tokens":3007}},"tokens_in":636,"tokens_out":3121,"duration_ms":21421,"temperature":1.0,"reasoning_tokens":3007,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T11:36:27.499473+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the straight-waveguide propagation loss directly by cut-back: compare end-to-end transmission through PPLNOI ridge waveguides of different lengths and see whether the extracted loss matches 0.106 dB/cm, and repeat the SHG slope measurement with independently calibrated facet coupling.","supporting_citations":[{"cited_title":"High-efficiency nonlinear frequency conversion enabled by optimizing the ferroelectric domain structure in x-cut LNOI ridge waveguide,","cited_arxiv_id":null,"evidence_quote":"Provides the previous best PPLNOI propagation-loss baseline of 0.23 dB/cm before poling and 0.25 dB/cm after, which this work claims to improve twofold."},{"cited_title":"Long Low-Loss-Litium Niobate on Insulator Waveguides with Sub-Nanometer Surface Roughness,","cited_arxiv_id":null,"evidence_quote":"Introduces the PLACE process showing sub-nanometer roughness in LNOI waveguides, the fabrication method this work adapts."},{"cited_title":"Ultra-broadband and low-loss edge coupler for highly efficient second harmonic generation in thin-film lithium niobate,","cited_arxiv_id":null,"evidence_quote":"Reports the 814%/W SHG conversion efficiency in e-beam/dry-etched PPLNOI, the record this work compares against."},{"cited_title":"Ultrahigh-efficiency wavelength conversion in nanophotonic periodically poled lithium niobate waveguides,","cited_arxiv_id":null,"evidence_quote":"Demonstrates ultrahigh-efficiency wavelength conversion in nanophotonic PPLN waveguides, establishing the efficiency class this work aims to approach."}],"review_version":1}